ad-creative
Warn
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill metadata and instructions point to an external, third-party GitHub repository for its implementation code. The instructions explicitly advise the user to install the "upstream bundle" from this source into the agent's skills directory, introducing a supply-chain risk from an untrusted source.
- URL: https://github.com/coreyhaines31/marketingskills
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and iterate upon user-provided ad creative text, which is an untrusted data source that could contain malicious instructions.
- Ingestion points: User-provided ad headlines, descriptions, and primary text (SKILL.md).
- Boundary markers: Absent.
- Capability inventory: None (the provided file is metadata-only and lacks internal script logic).
- Sanitization: Absent.
Audit Metadata